mem-search

Search rad-mem's persistent cross-session memory database for past work.

7|4|Updated Nov 27, 2025
One-click install
npx skills add https://github.com/thedotmack/rad-mem --skill mem-search
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mem-search
Source: https://github.com/thedotmack/rad-mem/tree/main/plugin/skills/mem-search
Command: npx skills add https://github.com/thedotmack/rad-mem --skill mem-search

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires node.js, pm2, sqlite3, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Claude Code sessions often hit context limits, making it hard to remember past work or project history. Manually sifting through old transcripts is time-consuming and inefficient. This Skill provides persistent memory and efficient retrieval, ensuring Claude always remembers what you've done.

Core Features & Use Cases

  • Persistent Cross-Session Memory: Automatically captures tool usage observations and generates semantic summaries, making them available for future sessions.
  • Skill-Based Search: Query your project history with natural language, saving tokens compared to manual context priming.
  • Progressive Disclosure: Retrieve layered memory (index first, then full details) to optimize token usage and focus on relevant information.
  • Use Case: Ask "What bugs did we fix last session?" or "How did we implement authentication?" to instantly retrieve relevant past work, code changes, and decisions without manually searching through logs.

Quick Start

Use the mem-search skill to find all bug fixes from last week.

Frequently Asked Questions about mem-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I search my past work across multiple Claude Code sessions?

Use mem-search to query your persistent cross-session memory database with natural language. Search returns indexed results showing relevant past work, decisions, and code changes from earlier sessions without manual transcript sifting.

Can I retrieve specific solutions from previous sessions without reloading full context?

Yes. mem-search uses a 3-step workflow: search your memory index, review results, then fetch full details only for selected IDs. This progressive disclosure saves tokens by fetching summaries first, then complete information only when needed.

How does mem-search help when I hit context limits in long projects?

Instead of manually priming context or searching logs, mem-search automatically captures and indexes past tool usage and semantic summaries. Query it to surface relevant prior work, reducing token overhead while keeping project history accessible across sessions.

What dependencies does mem-search require to function?

mem-search runs on Node.js, uses PM2 for process management, and SQLite3 as its persistent memory database. These dependencies enable cross-session storage and efficient querying of your work history.

Can mem-search work with automation and project management workflows?

Yes. mem-search integrates into automation and project management contexts by surfacing historical solutions, bug fixes, and implementation patterns. Query it to recall how you solved similar problems or what tasks ran in previous sessions.

What's the best way to structure queries when searching memory for past implementation details?

Use natural language queries like 'What bugs did we fix last session?' or 'How did we implement authentication?' mem-search interprets semantic intent, returning indexed summaries first so you can select relevant IDs to fetch full details.